An assistant professor at the University of Michigan School of Information, Dr. Julie Hui’s research seeks to understand and reduce barriers to workforce development. She was recently honored with a Provost’s Teaching Innovation Prize for her work with Lettersmith, a software helping students with professional writing. We chatted with Dr. Hui about Lettersmith, barriers to workplace equity, and how to expand accessibility in tech.

Hi Julie! Can you tell us about Lettersmith?
Lettersmith is a collaboration with the university’s Center for Academic Innovation. It’s a tool that guides students through the process of learning structured writing, specifically professional writing. This includes emails to potential employers asking for informational interviews, as well as examples of cover letters and college applications. The software offers free templates for students to use with built-in suggestions and writing advice.

What was the gap or need you noticed that you wanted Lettersmith to fill?
I initially developed Lettersmith during my PhD at Northwestern. I was working closely on a few community-based projects and noticed many teams had trouble communicating with their clients in basic ways; things like introducing themselves, describing the project, and setting agendas were difficult. I was thinking it might be useful to develop software to help with professional communication skills.

We prototyped a few different ideas and at first, the idea was a format similar to Mad Libs – the email is mostly written out for the user, and they fill in certain information. While easy to use, we got pushback because students felt like it took over their personal voice. So we redesigned Lettersmith in a way that provides guidance on best communication practices, but is ultimately up to the students to write.

I also developed Lettersmith because I personally feel like I got to where I am today by networking. Luckily, I had access to people who could read my messages and offer advice on my writing, but I know that not everyone has access to that type of mentorship. We were hoping this tool could at least provide some basic support and guidance for these conversations. While we’ve mostly focused on launching in the United States, we’d eventually like to expand internationally. The software has been very useful for non-native English speakers because, for some, it may be the first time they’re writing in a professional context in English – those expectations aren’t clearly stated anywhere.

It’s great you noticed a barrier to workplace equity and created a tool to expand accessibility. How have you navigated the challenges that come along with developing software?
Providing guidance on professional communication is important. At the same time, we want to make sure we’re not prescribing just one way to write. We know there are many different ways of writing and styles of communication. We thought about that when designing Lettersmith; you can provide many types of examples, so one user’s cover letter will look completely different from another. It’s important that recruiters and employers expand how they think about what professional communication looks like. Figuring out how to get people in power to rethink what professional communication looks like is a whole other challenge.

I’m glad you mentioned that because what we’ve defined as “professional” and “unprofessional” can be such coded language. How has Lettersmith engaged with that reality?
On a basic level, there are certain things many people agree to include if you’re networking: introducing yourself, sharing your interests and being clear about what you’re asking for. But in terms of cultural differences, those are much harder to tackle – what’s professional in one country can be very different from what’s professional in another. Within Lettersmith, we want to make it clear that one guide is not more professional than the other. We hope to make students aware of what these cultural differences are and they can decide whether or not they want to follow those.

What has the response been like to Lettersmith?
We’ve found that Lettersmith has improved the quality of users’ writing, as evaluated by professional career counselors. By quality, we mean that students are more likely to introduce themselves, describe their interests and the project, and incorporate key pieces of information. We have also seen improvements in student confidence, which we measure by how likely they are to send messages at all. We noticed that students would draft an email and it would sit in their inbox forever – writing professional messages can be really scary for students. We’re hoping that Lettersmith gets students to the point of actually sending it, and our research supports this! Lettersmith makes expectations of communication more transparent.

This concept of transparency has been shown to support equity in education when expectations can be siloed within certain networks. If we clarify what is expected within assignments or professional communication, students find it a lot easier to meet those goals. Interestingly, we also saw that using Lettersmith forced instructors to reflect on what they wanted to teach with professional writing. I assumed that instructors had a clear understanding of what they wanted to share with professional writing, but that wasn’t always the case. We’ve seen that Lettersmith has allowed instructors to come up with clear rubrics and expectations.

In addition to Lettersmith, you’ve worked as a theorist on socio-technical systems and human-computer interaction; I’m curious about your thoughts on AI and what you want to see implemented as it continues to develop?
There are definitely pros and cons with AI – and ChatGPT specifically. As professors, we’ve been having conversations about how to teach students with this new software. We want to acknowledge its existence, and try to teach responsible use. In an educational context, we hope students understand that simply copying and pasting without understanding the material is doing a disservice to their knowledge.

At the same time, if students don’t have access to mentors who can help them with skills like professional communication, AI could be a useful way to get feedback on drafts or emails. I think that’s perfectly reasonable, and expands access to resources and models students may not have seen before.

There’s a lot of conversation around how being able to use ChatGPT effectively is based on digital literacy levels. Not everyone can come up with the perfect search query to find the information they’re searching for – that takes background knowledge. I think some people believe that ChatGPT is going to improve access to information, but I would say that requires knowing how to search and what words to use. Like many other forms of technology, ChatGPT and AI in general will ultimately benefit those with higher writing and communication skills already. It’s important to acknowledge this reality.

That’s an important point in AI discourse – ChatGPT continues to benefit those who are already digitally literate and have access to technology. From your experience working in tech and information systems, how can we make these fields more accessible for those from marginalized communities?
This is a huge topic. I’ve noticed the importance of having opportunities to work on personally meaningful or societally impactful projects. While learning engineering skills, we need to be asking questions like: how could this influence the world? What could the negative impacts of this project be? Students are so aware of the adverse effects of technology, and they don’t want that to be a taboo topic in the classroom. Bringing this conversation into the curriculum both prepares students to be better technologists and engineers and brings in a more diverse set of students.

I’ve been thinking about a more care-based approach to instruction. In a lot of technical fields, the thought is that you learn “in the fire,” but we’ve found this isn’t inclusive. Being a tough instructor can be helpful but studies have shown that it needs to be in combination with showing you care about the students. If you show that you personally, truly care about them and you have high expectations, that’s a great combination. Lastly, to expand inclusivity within tech we need to make opportunities generally more accessible to students and show them that there are different ways to get help and resources.